Data Import
Import product catalogues, transactions, season tags, and outfit rules
Data Import is where your brand DNA journey begins. The platform analyses your product data to extract brand fingerprints, so the quality and completeness of your imports directly affects the quality of every insight downstream. This page handles product data, transaction data (for customer DNA), and context rules.
Product data should include product attributes labelled according to the Brand DNA ontology — 40 core attributes organised across 5 domains (Colour & Pattern, Shape & Structure, Material & Texture, Style & Identity, Commercial & Lifecycle). The column mapping tool helps translate your data's column names to the ontology's attribute names, either through auto-detection (fuzzy matching) or manual mapping.
The ontology browser shows the full attribute hierarchy: Domain → Resolution Level (MACRO/MESO/MICRO) → Attribute → Allowed Values. Use this to understand what data the platform expects and how your data maps to it. Attributes you don't have data for will simply show as 'no data' in the spectrogram — partial data is perfectly valid.
Transaction data (optional) enables customer DNA analysis. Upload purchase records linking customers to products, and the platform will compute per-customer attribute distributions — revealing what customers actually buy versus what you offer. Without transaction data, the customer alignment page uses product-based archetypes as a proxy.
Import validation checks your data against the ontology and flags issues: errors (blocking — e.g., missing required columns), warnings (non-blocking — e.g., unrecognised attribute values), and info messages (suggestions — e.g., similar column names that might be typos). The validation report helps you clean data before it affects analysis.
After every import, the system automatically recomputes all distributions and derived metrics. This includes brand DNA distributions, co-occurrence lifts, category consistency scores, and (if seasonal data is present) drift metrics. The recomputation typically takes a few seconds for datasets under 10,000 products.
Tips
- Use the column mapping tool when your CSV headers don't exactly match the ontology attribute names. The auto-mapper catches common variations (e.g., 'Color' → 'Predominant Colour').
- Start with product data — it's the foundation for everything. Transaction data and context rules are enhancements that unlock additional features.
- The sample data generator is useful for demos and testing. It creates synthetic but realistic product data across multiple brands.
- Check the import history table for past imports. If data quality seems off, you can review validation reports from previous uploads.
- When importing seasonal data, use consistent season identifiers (e.g., 'AW24', 'SS25') across imports so the drift analysis can correctly compare seasons.
- Larger datasets (10,000+ products) provide more statistically robust distributions. Small datasets (<100 products per brand) may show noisy patterns.
Enables: Brand DNA spectrogram, signal strength analysis, category consistency, and brand comparison
Sample file